Axial Capacity of Piles in Clay | Foundation Engineering - Civil Engineering (CE) PDF Download

 Page 1


1. Introduction
THERE ARE many areas of geotechnical
engineering where design is based largely
on empirical methods that have evolved
from previous experience. Such approa-
ches may work well in familiar situations,
but there will inevitably be uncertainty
when novel structures or new types of
soil deposits are encountered. Pile founda-
tions have been designed more or less
successfully, using an empirical approach,
for a large number of years. However, the
advent of the offshore oil industry has
necessitated the design of piles with ca-
pacity of an order of magnitude higher
than onshore piles, and in soils which are
in many instances different from onshore
deposits. This has provided considerable
impetus to improving the understanding
of how driven piles develop their capacity,
in order to provide a more rational basis
for pile design.
Nearly all the working load capacity of
driven piles comes from the shaft, and this
Paper is restricted to discussion of devel-
opments in the prediction of the shaft ca-
pacity of piles driven into clay. Until rela-
tively recently, the prevalent design phil-
osophy was that some bond or adhesion
existed between the pile shaft and the soil.
It was natural to correlate the strength of
this bond with the undrained shear stren-
gth of the soil. This led to what is com-
monly referred to as the 'total stress met-
hod'or
predicting pile capacity. The limit-
ing skin friction, —,„,on the pile shaft is
expressed as a proportion of the in-situ
shear strength of the soil,
cu,
as
<a
=
"cue
Experience obtained from onshore pile
tests has enabled the parameter x to be
correlated with the type and strength of
the soil (e.g. Tomlinson 1980).
The first advances from this standpoint
were due to Chandler (1968) and Burland
(1973) who argued that soil behaviour is
controlled by effective stresses and that
the shaft capacity of a pile should be view-
ed in the light of a limiting frictional stress
developed between the pile and the soil.
The new 'effective stress'pproach sought
to relate the skin friction to the in-situ
effective stress state. Although it was rec-
ognised that the limiting
sk'n
friction would
depend primarily on the horizontal effec-
tive stress, this quantity is difficult to esti-
mate with confidence. The most accurately
determinable effective stress in the soil is
the effective overburden pressure,
rre,
'.
Consequently, the effect've stress a p-
proach has led to the skin friction being
expressed as some fraction,
P,
of the effec-
«University of Cambridge
liuniversity of Oxford
This Paper will form the bas's of a presentation
by the authors to an Informal Discussion meet-
ing of the British Geotechnical Society on Dec. t
this year at the Institution of Civil Engineers,
Westminster, London.
tive overburden pressure according to
Ta P(rte (2)
The parameter
P
thus reflects not only
the friction angle between the pile and the
soil, but also the ratio of the horizontal
and vertical effective stresses. Burland
(1973) suggested that this ratio could be
taken as equal to the at-rest earth pres-
sure coefficient,
K,, acknowledging that
this would be conservative for driven piles,
where the act of installing the pile may
locally increase the horizontal stress in the
soil.
Although derived from different stand-
points, for the designer there is no particu-
lar merit to the expression in eqn. 2 com-
pared with that in eqn. 1. In both cases,
the skin friction is directly and simply re-
lated to measurable properties of the soil
which reflect its strength prior to install-
ing a pile. But of course, the advantage
of the effective stress approach is that it
opens the door to considering, in detail,
precisely what changes in total and effec-
tive stresses take place due to the irstal-
lation of the pile and due to its subse-
quent loading.
This Paper outlines some of the more
recent developments that have occurred
in the effective stress approach to pile de-
sign. In particular, attention is concentra-
ted on three main events in the history of
a driven pile. These are
(i) pile installation
(ii
)
consolidation, as excess pore pres-
sures generated in (i) dissipate, and
(iii) pile loading.
Much of the work described herein was
instigated by the ESACC project, organised
by Amoco Production Company and fund-
ed by 17 assorted organisations. A sum-
mary of this project may be found in Kraft
et al (1980) and a review of the effective
stress methods of design that emerged
from the project is given by Kraft (1982).
One of the chief limitations to formu-
lating new design rules has been the pauc-
ity of reliable data on the shaft capacity
of piles, particularly in the types of soil
encountered offshore. In the absence of
such data, some of the discussion in the
present Paper is necessarily of a specula-
tive nature; the Paper should be viewed
as an attempt to draw preliminary design
guidelines from research which is still con-
tinuing and has some way to go before
more definitive statements may be made.
2. Changes in stresses
In discussing the stress changes which
occur in the soil, it is convenient to adopt
cylindrical polar co-ordinates r, 6'nd
z,
with the z axis directed down the axis of
the pile as shown in Fig. 1. Discussion will
be confined to cylindrical piles but the
principles developed are also applicable to
piles of other cross-sectional shapes or to
tapered piles. Prior to pile installation, the
stresses acting on a typical element of
soil are the 'at rest'tresses 0.,' —
<r„o'nd
rr
'
tran
'
K rr +', as shown in Fig.
1. In accordance with common practice, it
is assumed that the horizontal stresses
are the same in any direction. The changes
in the stresses during and after pile instal-
lation are now discussed.
2.1 Pile installation
The process of installing a displacement
pile into clay is one which is not easily
modelled analytically. Fig. 2 shows sche-
matically the main soil movements which
occur. The most common idealisation of
these movements has been that of cylin-
drical cavity expansion (e.g. Soderberg,
1962). The idealisation neglects local ef-
fects such as heave of the ground surface
around the pile and does not
attempt
to
<')
~0
r
Pile
Soil element
Fig. 1. Co-ordinate system and initial
stress state in the soil
Pile~— Surface
heave
~)Z,
I
I
Zone of intensely~',
remoulded so
"l
Radial movement
of soil
Spherical
pressure bulb
Fig. 2. Soil movements due to pile instal-
lationn
Page 2


1. Introduction
THERE ARE many areas of geotechnical
engineering where design is based largely
on empirical methods that have evolved
from previous experience. Such approa-
ches may work well in familiar situations,
but there will inevitably be uncertainty
when novel structures or new types of
soil deposits are encountered. Pile founda-
tions have been designed more or less
successfully, using an empirical approach,
for a large number of years. However, the
advent of the offshore oil industry has
necessitated the design of piles with ca-
pacity of an order of magnitude higher
than onshore piles, and in soils which are
in many instances different from onshore
deposits. This has provided considerable
impetus to improving the understanding
of how driven piles develop their capacity,
in order to provide a more rational basis
for pile design.
Nearly all the working load capacity of
driven piles comes from the shaft, and this
Paper is restricted to discussion of devel-
opments in the prediction of the shaft ca-
pacity of piles driven into clay. Until rela-
tively recently, the prevalent design phil-
osophy was that some bond or adhesion
existed between the pile shaft and the soil.
It was natural to correlate the strength of
this bond with the undrained shear stren-
gth of the soil. This led to what is com-
monly referred to as the 'total stress met-
hod'or
predicting pile capacity. The limit-
ing skin friction, —,„,on the pile shaft is
expressed as a proportion of the in-situ
shear strength of the soil,
cu,
as
<a
=
"cue
Experience obtained from onshore pile
tests has enabled the parameter x to be
correlated with the type and strength of
the soil (e.g. Tomlinson 1980).
The first advances from this standpoint
were due to Chandler (1968) and Burland
(1973) who argued that soil behaviour is
controlled by effective stresses and that
the shaft capacity of a pile should be view-
ed in the light of a limiting frictional stress
developed between the pile and the soil.
The new 'effective stress'pproach sought
to relate the skin friction to the in-situ
effective stress state. Although it was rec-
ognised that the limiting
sk'n
friction would
depend primarily on the horizontal effec-
tive stress, this quantity is difficult to esti-
mate with confidence. The most accurately
determinable effective stress in the soil is
the effective overburden pressure,
rre,
'.
Consequently, the effect've stress a p-
proach has led to the skin friction being
expressed as some fraction,
P,
of the effec-
«University of Cambridge
liuniversity of Oxford
This Paper will form the bas's of a presentation
by the authors to an Informal Discussion meet-
ing of the British Geotechnical Society on Dec. t
this year at the Institution of Civil Engineers,
Westminster, London.
tive overburden pressure according to
Ta P(rte (2)
The parameter
P
thus reflects not only
the friction angle between the pile and the
soil, but also the ratio of the horizontal
and vertical effective stresses. Burland
(1973) suggested that this ratio could be
taken as equal to the at-rest earth pres-
sure coefficient,
K,, acknowledging that
this would be conservative for driven piles,
where the act of installing the pile may
locally increase the horizontal stress in the
soil.
Although derived from different stand-
points, for the designer there is no particu-
lar merit to the expression in eqn. 2 com-
pared with that in eqn. 1. In both cases,
the skin friction is directly and simply re-
lated to measurable properties of the soil
which reflect its strength prior to install-
ing a pile. But of course, the advantage
of the effective stress approach is that it
opens the door to considering, in detail,
precisely what changes in total and effec-
tive stresses take place due to the irstal-
lation of the pile and due to its subse-
quent loading.
This Paper outlines some of the more
recent developments that have occurred
in the effective stress approach to pile de-
sign. In particular, attention is concentra-
ted on three main events in the history of
a driven pile. These are
(i) pile installation
(ii
)
consolidation, as excess pore pres-
sures generated in (i) dissipate, and
(iii) pile loading.
Much of the work described herein was
instigated by the ESACC project, organised
by Amoco Production Company and fund-
ed by 17 assorted organisations. A sum-
mary of this project may be found in Kraft
et al (1980) and a review of the effective
stress methods of design that emerged
from the project is given by Kraft (1982).
One of the chief limitations to formu-
lating new design rules has been the pauc-
ity of reliable data on the shaft capacity
of piles, particularly in the types of soil
encountered offshore. In the absence of
such data, some of the discussion in the
present Paper is necessarily of a specula-
tive nature; the Paper should be viewed
as an attempt to draw preliminary design
guidelines from research which is still con-
tinuing and has some way to go before
more definitive statements may be made.
2. Changes in stresses
In discussing the stress changes which
occur in the soil, it is convenient to adopt
cylindrical polar co-ordinates r, 6'nd
z,
with the z axis directed down the axis of
the pile as shown in Fig. 1. Discussion will
be confined to cylindrical piles but the
principles developed are also applicable to
piles of other cross-sectional shapes or to
tapered piles. Prior to pile installation, the
stresses acting on a typical element of
soil are the 'at rest'tresses 0.,' —
<r„o'nd
rr
'
tran
'
K rr +', as shown in Fig.
1. In accordance with common practice, it
is assumed that the horizontal stresses
are the same in any direction. The changes
in the stresses during and after pile instal-
lation are now discussed.
2.1 Pile installation
The process of installing a displacement
pile into clay is one which is not easily
modelled analytically. Fig. 2 shows sche-
matically the main soil movements which
occur. The most common idealisation of
these movements has been that of cylin-
drical cavity expansion (e.g. Soderberg,
1962). The idealisation neglects local ef-
fects such as heave of the ground surface
around the pile and does not
attempt
to
<')
~0
r
Pile
Soil element
Fig. 1. Co-ordinate system and initial
stress state in the soil
Pile~— Surface
heave
~)Z,
I
I
Zone of intensely~',
remoulded so
"l
Radial movement
of soil
Spherical
pressure bulb
Fig. 2. Soil movements due to pile instal-
lationn
o,'-
1.5cuo oz 3cuo
or 2.5cuo u,
—
5cu
(ai Stresses on element adjacent to pile (a) Stresses on element adjacent to pile
Stress
cuo
Critical state
region
I
I
Stress
uo
4-
OCR=1
Ko
=
0.55
OCR=1
Ko
=
0.55
0
1 2
(bi Stress distribution around pile
10
r I II> I
100
I'/I'
1
I I I I I I > I
10
(bi Stress distribution around pile
I I I I I I
100
llr
Fig. 3. Theoretical stress state afterinstallation of pile in normally
consolidated soil (from Randolph er ak 1979b)
Fig. 5. Theoretical stress state after consolidation around pile in
normally consolidated soil (from Randolph er af, 1979b)
~
doi
3
lj
2-
(su/ovo~u
~+
1'
1
I
4 8
OCR
Fig. 4. Excess pore pressures generated
c/ose to a model jacked pile during instal-
latiOn (from Francescon, 1982)
model the precise detail of the soil move-
ment near the advancing pile tip. However,
it has been shown (Randolph et al, 1979a;
Steenfelt et al, 1981) that, over most of
the pile length, the radial soil displace-
ments due to installation of a displace-
ment pile are described well by idealising
the installation as the expansion of a cy-
lindrical cavity under conditions of plane
strain. This has enabled quantitative esti-
mates to be made of the stress changes
which occur in the soil around the pile
(Carter et al, 1979a; Randolph et al, 1979b) .
Idealisation of pile installation by cavity
expansion assumes that a cylindrical hole
is created, from zero initial radius, by radial
movement of the soil. The idealisation in-
volves a similar strain path to that follow-
ed in a pressuremeter test, and the total
stresses acting on the pile after cavity
expansion correspond to the limit pressure
achieved in such a test. Since the cavity
expansion starts from zero radius, the
shear strains in the soil are essentially
infinite adjacent to the pile wall, and are
such as to bring the soil to a state of fail-
ure out to a radius of 7-10 times the pile
radius, r,. Within this zone, the soil will
be remoulded to a greater or lesser extent.
The concepts embodied in critical state
soil mechanics indicate that soil sheared
to failure under undrained conditions will
tend to a unique effective stress state for
a given initial water content. Thus the
effective stress state in the failed zone of
soil around the pile will be uniform at any
particular depth. The total stress level will
vary with the radial distance from the pile
and the different changes in the total and
effective stress levels within any soil ele-
ment will give rise to excess pore pres-
sures. Typically, these may be as high as
4-6 times the in-situ shear strength close
to piles driven into normally or lightly
overconsolidated soils.
In the zone of soil sheared to critical
state conditions, where the effect've stress
conditions are uniform the shear strength
of the soil is fully mobilised in the hori-
zontal plane such that n-„'
n.fr
'
2c„,.
The equation of radial equilibrium then
implies that, within this zone, the radial
total stress, and hence the excess pore
pressures, should vary linearly with the
logarithm of the radial distance from the
pile axis, w'ith a gradient of 2c„,. Randolph
et al (1979b) have used the modified Cam
clay model (based on that described by
Roscoe 8f Burland, 1968) to provide quan-
titative estimates of the stresses after
cavity expansion. These are shown in Fig.
3 for soil with a friction angle of y'
30'.
5u
=
4c„,
—
3p'3)
where Ap's the change in mean effective
stress due to shearing the soil to a critical
state condition. For normally consolidated
soil, Ap'ill be negative and will typically
be of magnitude 1-1.5c„,.At increasing
values of overconsolidation ratio (OCR).
Ap'ill increase, becoming positive for
values of OCR greater than 2-3. Thus for
soil
'wet'f
critical (in the terminology
of Schofield 8s Wroth, 1968) the excess
pore pressures given by eqn. 3 will be
greater than
4cu,,
while for heavily over-
The numerical analyses showed little
variation of the size of the predicted ex-
cess pore pressures with the degree of
overconsolidation of the soil. However,
this result is questionable in view of the
greater tendency of heavily overconsolida-
ted soils to dilate. The excess pore pres-
sures arise partly from the increase in
mean total stress that accompanies cavity
expansion, an() partly as a result of chan-
ges in the mean effective stress as the soil
is sheared and remoulded. For normally or
lightly overconsolidated clay, the mean
effective stress will decrease during un-
drained shearing to failure, while for heav-
ily overconsolidated clay, there will be
an increase in the mean effective stress
(as the soil attempts to dilate). It may
therefore be expected that higher excess
pore pressures
w'll
be generated by pile
installation in lightly overconsolidated clay
than in heavily overconsolidated clay.
Randolph et al (1979b) suggest an ap-
proximate expression for estimating the
excess pore pressures generated adjacent
to a driven pile, which is
Page 3


1. Introduction
THERE ARE many areas of geotechnical
engineering where design is based largely
on empirical methods that have evolved
from previous experience. Such approa-
ches may work well in familiar situations,
but there will inevitably be uncertainty
when novel structures or new types of
soil deposits are encountered. Pile founda-
tions have been designed more or less
successfully, using an empirical approach,
for a large number of years. However, the
advent of the offshore oil industry has
necessitated the design of piles with ca-
pacity of an order of magnitude higher
than onshore piles, and in soils which are
in many instances different from onshore
deposits. This has provided considerable
impetus to improving the understanding
of how driven piles develop their capacity,
in order to provide a more rational basis
for pile design.
Nearly all the working load capacity of
driven piles comes from the shaft, and this
Paper is restricted to discussion of devel-
opments in the prediction of the shaft ca-
pacity of piles driven into clay. Until rela-
tively recently, the prevalent design phil-
osophy was that some bond or adhesion
existed between the pile shaft and the soil.
It was natural to correlate the strength of
this bond with the undrained shear stren-
gth of the soil. This led to what is com-
monly referred to as the 'total stress met-
hod'or
predicting pile capacity. The limit-
ing skin friction, —,„,on the pile shaft is
expressed as a proportion of the in-situ
shear strength of the soil,
cu,
as
<a
=
"cue
Experience obtained from onshore pile
tests has enabled the parameter x to be
correlated with the type and strength of
the soil (e.g. Tomlinson 1980).
The first advances from this standpoint
were due to Chandler (1968) and Burland
(1973) who argued that soil behaviour is
controlled by effective stresses and that
the shaft capacity of a pile should be view-
ed in the light of a limiting frictional stress
developed between the pile and the soil.
The new 'effective stress'pproach sought
to relate the skin friction to the in-situ
effective stress state. Although it was rec-
ognised that the limiting
sk'n
friction would
depend primarily on the horizontal effec-
tive stress, this quantity is difficult to esti-
mate with confidence. The most accurately
determinable effective stress in the soil is
the effective overburden pressure,
rre,
'.
Consequently, the effect've stress a p-
proach has led to the skin friction being
expressed as some fraction,
P,
of the effec-
«University of Cambridge
liuniversity of Oxford
This Paper will form the bas's of a presentation
by the authors to an Informal Discussion meet-
ing of the British Geotechnical Society on Dec. t
this year at the Institution of Civil Engineers,
Westminster, London.
tive overburden pressure according to
Ta P(rte (2)
The parameter
P
thus reflects not only
the friction angle between the pile and the
soil, but also the ratio of the horizontal
and vertical effective stresses. Burland
(1973) suggested that this ratio could be
taken as equal to the at-rest earth pres-
sure coefficient,
K,, acknowledging that
this would be conservative for driven piles,
where the act of installing the pile may
locally increase the horizontal stress in the
soil.
Although derived from different stand-
points, for the designer there is no particu-
lar merit to the expression in eqn. 2 com-
pared with that in eqn. 1. In both cases,
the skin friction is directly and simply re-
lated to measurable properties of the soil
which reflect its strength prior to install-
ing a pile. But of course, the advantage
of the effective stress approach is that it
opens the door to considering, in detail,
precisely what changes in total and effec-
tive stresses take place due to the irstal-
lation of the pile and due to its subse-
quent loading.
This Paper outlines some of the more
recent developments that have occurred
in the effective stress approach to pile de-
sign. In particular, attention is concentra-
ted on three main events in the history of
a driven pile. These are
(i) pile installation
(ii
)
consolidation, as excess pore pres-
sures generated in (i) dissipate, and
(iii) pile loading.
Much of the work described herein was
instigated by the ESACC project, organised
by Amoco Production Company and fund-
ed by 17 assorted organisations. A sum-
mary of this project may be found in Kraft
et al (1980) and a review of the effective
stress methods of design that emerged
from the project is given by Kraft (1982).
One of the chief limitations to formu-
lating new design rules has been the pauc-
ity of reliable data on the shaft capacity
of piles, particularly in the types of soil
encountered offshore. In the absence of
such data, some of the discussion in the
present Paper is necessarily of a specula-
tive nature; the Paper should be viewed
as an attempt to draw preliminary design
guidelines from research which is still con-
tinuing and has some way to go before
more definitive statements may be made.
2. Changes in stresses
In discussing the stress changes which
occur in the soil, it is convenient to adopt
cylindrical polar co-ordinates r, 6'nd
z,
with the z axis directed down the axis of
the pile as shown in Fig. 1. Discussion will
be confined to cylindrical piles but the
principles developed are also applicable to
piles of other cross-sectional shapes or to
tapered piles. Prior to pile installation, the
stresses acting on a typical element of
soil are the 'at rest'tresses 0.,' —
<r„o'nd
rr
'
tran
'
K rr +', as shown in Fig.
1. In accordance with common practice, it
is assumed that the horizontal stresses
are the same in any direction. The changes
in the stresses during and after pile instal-
lation are now discussed.
2.1 Pile installation
The process of installing a displacement
pile into clay is one which is not easily
modelled analytically. Fig. 2 shows sche-
matically the main soil movements which
occur. The most common idealisation of
these movements has been that of cylin-
drical cavity expansion (e.g. Soderberg,
1962). The idealisation neglects local ef-
fects such as heave of the ground surface
around the pile and does not
attempt
to
<')
~0
r
Pile
Soil element
Fig. 1. Co-ordinate system and initial
stress state in the soil
Pile~— Surface
heave
~)Z,
I
I
Zone of intensely~',
remoulded so
"l
Radial movement
of soil
Spherical
pressure bulb
Fig. 2. Soil movements due to pile instal-
lationn
o,'-
1.5cuo oz 3cuo
or 2.5cuo u,
—
5cu
(ai Stresses on element adjacent to pile (a) Stresses on element adjacent to pile
Stress
cuo
Critical state
region
I
I
Stress
uo
4-
OCR=1
Ko
=
0.55
OCR=1
Ko
=
0.55
0
1 2
(bi Stress distribution around pile
10
r I II> I
100
I'/I'
1
I I I I I I > I
10
(bi Stress distribution around pile
I I I I I I
100
llr
Fig. 3. Theoretical stress state afterinstallation of pile in normally
consolidated soil (from Randolph er ak 1979b)
Fig. 5. Theoretical stress state after consolidation around pile in
normally consolidated soil (from Randolph er af, 1979b)
~
doi
3
lj
2-
(su/ovo~u
~+
1'
1
I
4 8
OCR
Fig. 4. Excess pore pressures generated
c/ose to a model jacked pile during instal-
latiOn (from Francescon, 1982)
model the precise detail of the soil move-
ment near the advancing pile tip. However,
it has been shown (Randolph et al, 1979a;
Steenfelt et al, 1981) that, over most of
the pile length, the radial soil displace-
ments due to installation of a displace-
ment pile are described well by idealising
the installation as the expansion of a cy-
lindrical cavity under conditions of plane
strain. This has enabled quantitative esti-
mates to be made of the stress changes
which occur in the soil around the pile
(Carter et al, 1979a; Randolph et al, 1979b) .
Idealisation of pile installation by cavity
expansion assumes that a cylindrical hole
is created, from zero initial radius, by radial
movement of the soil. The idealisation in-
volves a similar strain path to that follow-
ed in a pressuremeter test, and the total
stresses acting on the pile after cavity
expansion correspond to the limit pressure
achieved in such a test. Since the cavity
expansion starts from zero radius, the
shear strains in the soil are essentially
infinite adjacent to the pile wall, and are
such as to bring the soil to a state of fail-
ure out to a radius of 7-10 times the pile
radius, r,. Within this zone, the soil will
be remoulded to a greater or lesser extent.
The concepts embodied in critical state
soil mechanics indicate that soil sheared
to failure under undrained conditions will
tend to a unique effective stress state for
a given initial water content. Thus the
effective stress state in the failed zone of
soil around the pile will be uniform at any
particular depth. The total stress level will
vary with the radial distance from the pile
and the different changes in the total and
effective stress levels within any soil ele-
ment will give rise to excess pore pres-
sures. Typically, these may be as high as
4-6 times the in-situ shear strength close
to piles driven into normally or lightly
overconsolidated soils.
In the zone of soil sheared to critical
state conditions, where the effect've stress
conditions are uniform the shear strength
of the soil is fully mobilised in the hori-
zontal plane such that n-„'
n.fr
'
2c„,.
The equation of radial equilibrium then
implies that, within this zone, the radial
total stress, and hence the excess pore
pressures, should vary linearly with the
logarithm of the radial distance from the
pile axis, w'ith a gradient of 2c„,. Randolph
et al (1979b) have used the modified Cam
clay model (based on that described by
Roscoe 8f Burland, 1968) to provide quan-
titative estimates of the stresses after
cavity expansion. These are shown in Fig.
3 for soil with a friction angle of y'
30'.
5u
=
4c„,
—
3p'3)
where Ap's the change in mean effective
stress due to shearing the soil to a critical
state condition. For normally consolidated
soil, Ap'ill be negative and will typically
be of magnitude 1-1.5c„,.At increasing
values of overconsolidation ratio (OCR).
Ap'ill increase, becoming positive for
values of OCR greater than 2-3. Thus for
soil
'wet'f
critical (in the terminology
of Schofield 8s Wroth, 1968) the excess
pore pressures given by eqn. 3 will be
greater than
4cu,,
while for heavily over-
The numerical analyses showed little
variation of the size of the predicted ex-
cess pore pressures with the degree of
overconsolidation of the soil. However,
this result is questionable in view of the
greater tendency of heavily overconsolida-
ted soils to dilate. The excess pore pres-
sures arise partly from the increase in
mean total stress that accompanies cavity
expansion, an() partly as a result of chan-
ges in the mean effective stress as the soil
is sheared and remoulded. For normally or
lightly overconsolidated clay, the mean
effective stress will decrease during un-
drained shearing to failure, while for heav-
ily overconsolidated clay, there will be
an increase in the mean effective stress
(as the soil attempts to dilate). It may
therefore be expected that higher excess
pore pressures
w'll
be generated by pile
installation in lightly overconsolidated clay
than in heavily overconsolidated clay.
Randolph et al (1979b) suggest an ap-
proximate expression for estimating the
excess pore pressures generated adjacent
to a driven pile, which is
consolidated soils on the 'dry'ide of criti-
cal, <<p'ill be positive and excess pore
pressures of less than 4c„„are predicted.
This trend of decreasing excess pore pres-
sure with increasing OCR has been found
in model pile tests reported by
Frances-
con (1982). Fig. 4 shows some of his ex-
perimental results, where the excess pore
pressures generated at the pile shaft dur-
ing installation have been plotted against
the overconsolidation ratio of the clay.
The excess pore pressures have been nor-
malised by the initial undrained shear
strength,
c„„,
and also by the effective
overburden pressure, <re„'. Four test re-
sults are shown, corresponding to OCR of
1, 2, 4 and 8. The ratio
~u/ce,, decreases
from just under 5 at OCR
=
1 down to 2.6
at OCR
=
8. Correspondingly, the ratio
Au/<rr„'ncreases from unity at OCR
1, up to 2.7 at OCR =
8.
Although the simple model of cylindrical
cavity expansion has been justified on the
basis of the final radial movement of soil
around the pile, the stress changes will be
affected by the detailed strain path follow-
ed. A more detailed study of the stress
changes that occur has been made by
Levadoux (1980) by following the strain
path undergone by soil close to the ad-
vancing pile tip. The computations lead to
rather lower values of excess pore pres-
sures, particularly for heavily overconsoli-
dated soil.
2.2 Consolidation
The maximum hydraulic gradients exist-
ing immediately after installation of the
pile are primarily radial. At the pile wall,
the pile precludes straining of the soil, eit-
her along the pile length or around the
pile circumference, during consolidation.
Consequently, soil close to the pile under-
goes a form of one-dimensional consolida-
tion where the only non-zero strain is nor-
mal to the pile shaft (i.e. in the radial
direction). It is to be expected that the
radial effective stress will be the major
principal stress at the end of consolida-
tion. Finite element analyses have confirm-
ed this, and have shown that the vertical
and circumferential stresses at the end of
consolidation are approximately equal
(Randolph et al, 1979b). The finite element
analyses assumed that consolidation oc-
curs under plane strain conditions (with
no strain parallel to the pile axis, at any
radius). This assumption neglects the fact
that, in reality, there is a free ground sur-
face which may allow some straining of
the soil in the vertical direction.
For the assumption of plane strain, num-
erical predictions of the stress distribution
at the end of consolidation are as shown
in Fig, 5. Close to the pile, the radial and
vertical effective stresses are calculated
to be
rr5Ce<<<r3cec (4)
These values are reasonable for normal-
ly or lightly overconsolidated soil, where
c„„/<rr„'ay typically lie close to 0.3.
Thus the predicted final value of
<r
's
close to <rr„'. However, for heavily
overconsolidated clay, where
c„,/<rne may
be as high as 2 or
3, the predicted vertical
stre s appears unrealistically high, Free-
dom to strain in the vertical direction is
likely to give rise to rather lower values of
<rc
't
the end of the consolidation phase.
Of course, values of <r 'reater than the
effective overburden pressure may still de-
velop during consolidation, with vertical
equilibrium being maintained
by residual
shear stresses in the r:z plane. However,
during the working life of the pile, this
residual stress regime will alter and it
would seem reasonable to assume that o.
'ill
ultimately return to a value close to
the effective overburden pressure, n-„,'.
The consequences of this different be-
haviour, depending on the overconsolida-
tion ratio of the soil, are illustrated in Fig.
6 and 7. Fig. 6 shows stress paths obtained
by Randolph et al (1979b) for normally
consolidated clay. The paths are plotted
in q:p'nd V:p'pace, where p's
the
mean effective stress,
q
is the generalised
deviator stress given by qs =
0.5 [(<r
<ts')a + (<r.,' <r,')s +(o-s'-<T,')
'-'],
and
V is the specific volume of the soil. Start-
ing from point A, and ignoring ambient
pore pressures, the stress paths during
pile installation are shown by
AB'effec-
tive) and AB (total). During consolida-
tion, the effective and total stress states
converge, ending at point C. At this stage,
the water content of the soil has de-
creased and the soil strength will have in-
creased by 40
—
60%. This predicted be-
haviour is in good agreement with the
careful measurements of changes in soil
properties around piles driven in the field,
reported by Seed 8< Reese (1955). Wroth
et al (1979) report the results of finite
element analyses using the modified Cam-
clay model, showing how the increase in
shear strength varies with the consolida-
tion and strength properties of the soil.
Fig. 8 shows these results, with the
increase in shear strength,
c„„ /c„,,
(where c„represents the shear strength
of the soil after consolidation is complete)
being plotted against the Cam-clay para-
meter M. This parameter is related to the
friction angle <I>
'or
triaxial compression
by M
=
6sin<i<'/(3
—
sin<~').
As discussed earlier, the stress paths for
heavily overconsolidated soil may differ
from those for normally consolidated soil,
especially where straining is permitted in
the vertical direction during consolidation.
Fig. 7 indicates possible stress paths for
heavily overconsolidated soil. During in-
stallation, lower excess pore pressures are
generated, and the final effective stress
level after consolidation, is controlled by
the condition that <t-.'eturns to close to
<r„„'. The increase in the strength of the
soil during consolidation will be less than
for the case of a pile driven into normally
consolidated soil.
As consolidation proceeds, the effective
stresses adjacent to the pile will increase
Critical
state line
Total stress
paths
+B
Total stress
paths
P,P
P,P
Normal
consolidation line
Critical
state line
ormal
nsolidation line
Critical
state I <ne
Fig. 6. Stress paths for normally consolidated soil Fig. 7. Stress paths for heavily overconsolidated soil
P
Page 4


1. Introduction
THERE ARE many areas of geotechnical
engineering where design is based largely
on empirical methods that have evolved
from previous experience. Such approa-
ches may work well in familiar situations,
but there will inevitably be uncertainty
when novel structures or new types of
soil deposits are encountered. Pile founda-
tions have been designed more or less
successfully, using an empirical approach,
for a large number of years. However, the
advent of the offshore oil industry has
necessitated the design of piles with ca-
pacity of an order of magnitude higher
than onshore piles, and in soils which are
in many instances different from onshore
deposits. This has provided considerable
impetus to improving the understanding
of how driven piles develop their capacity,
in order to provide a more rational basis
for pile design.
Nearly all the working load capacity of
driven piles comes from the shaft, and this
Paper is restricted to discussion of devel-
opments in the prediction of the shaft ca-
pacity of piles driven into clay. Until rela-
tively recently, the prevalent design phil-
osophy was that some bond or adhesion
existed between the pile shaft and the soil.
It was natural to correlate the strength of
this bond with the undrained shear stren-
gth of the soil. This led to what is com-
monly referred to as the 'total stress met-
hod'or
predicting pile capacity. The limit-
ing skin friction, —,„,on the pile shaft is
expressed as a proportion of the in-situ
shear strength of the soil,
cu,
as
<a
=
"cue
Experience obtained from onshore pile
tests has enabled the parameter x to be
correlated with the type and strength of
the soil (e.g. Tomlinson 1980).
The first advances from this standpoint
were due to Chandler (1968) and Burland
(1973) who argued that soil behaviour is
controlled by effective stresses and that
the shaft capacity of a pile should be view-
ed in the light of a limiting frictional stress
developed between the pile and the soil.
The new 'effective stress'pproach sought
to relate the skin friction to the in-situ
effective stress state. Although it was rec-
ognised that the limiting
sk'n
friction would
depend primarily on the horizontal effec-
tive stress, this quantity is difficult to esti-
mate with confidence. The most accurately
determinable effective stress in the soil is
the effective overburden pressure,
rre,
'.
Consequently, the effect've stress a p-
proach has led to the skin friction being
expressed as some fraction,
P,
of the effec-
«University of Cambridge
liuniversity of Oxford
This Paper will form the bas's of a presentation
by the authors to an Informal Discussion meet-
ing of the British Geotechnical Society on Dec. t
this year at the Institution of Civil Engineers,
Westminster, London.
tive overburden pressure according to
Ta P(rte (2)
The parameter
P
thus reflects not only
the friction angle between the pile and the
soil, but also the ratio of the horizontal
and vertical effective stresses. Burland
(1973) suggested that this ratio could be
taken as equal to the at-rest earth pres-
sure coefficient,
K,, acknowledging that
this would be conservative for driven piles,
where the act of installing the pile may
locally increase the horizontal stress in the
soil.
Although derived from different stand-
points, for the designer there is no particu-
lar merit to the expression in eqn. 2 com-
pared with that in eqn. 1. In both cases,
the skin friction is directly and simply re-
lated to measurable properties of the soil
which reflect its strength prior to install-
ing a pile. But of course, the advantage
of the effective stress approach is that it
opens the door to considering, in detail,
precisely what changes in total and effec-
tive stresses take place due to the irstal-
lation of the pile and due to its subse-
quent loading.
This Paper outlines some of the more
recent developments that have occurred
in the effective stress approach to pile de-
sign. In particular, attention is concentra-
ted on three main events in the history of
a driven pile. These are
(i) pile installation
(ii
)
consolidation, as excess pore pres-
sures generated in (i) dissipate, and
(iii) pile loading.
Much of the work described herein was
instigated by the ESACC project, organised
by Amoco Production Company and fund-
ed by 17 assorted organisations. A sum-
mary of this project may be found in Kraft
et al (1980) and a review of the effective
stress methods of design that emerged
from the project is given by Kraft (1982).
One of the chief limitations to formu-
lating new design rules has been the pauc-
ity of reliable data on the shaft capacity
of piles, particularly in the types of soil
encountered offshore. In the absence of
such data, some of the discussion in the
present Paper is necessarily of a specula-
tive nature; the Paper should be viewed
as an attempt to draw preliminary design
guidelines from research which is still con-
tinuing and has some way to go before
more definitive statements may be made.
2. Changes in stresses
In discussing the stress changes which
occur in the soil, it is convenient to adopt
cylindrical polar co-ordinates r, 6'nd
z,
with the z axis directed down the axis of
the pile as shown in Fig. 1. Discussion will
be confined to cylindrical piles but the
principles developed are also applicable to
piles of other cross-sectional shapes or to
tapered piles. Prior to pile installation, the
stresses acting on a typical element of
soil are the 'at rest'tresses 0.,' —
<r„o'nd
rr
'
tran
'
K rr +', as shown in Fig.
1. In accordance with common practice, it
is assumed that the horizontal stresses
are the same in any direction. The changes
in the stresses during and after pile instal-
lation are now discussed.
2.1 Pile installation
The process of installing a displacement
pile into clay is one which is not easily
modelled analytically. Fig. 2 shows sche-
matically the main soil movements which
occur. The most common idealisation of
these movements has been that of cylin-
drical cavity expansion (e.g. Soderberg,
1962). The idealisation neglects local ef-
fects such as heave of the ground surface
around the pile and does not
attempt
to
<')
~0
r
Pile
Soil element
Fig. 1. Co-ordinate system and initial
stress state in the soil
Pile~— Surface
heave
~)Z,
I
I
Zone of intensely~',
remoulded so
"l
Radial movement
of soil
Spherical
pressure bulb
Fig. 2. Soil movements due to pile instal-
lationn
o,'-
1.5cuo oz 3cuo
or 2.5cuo u,
—
5cu
(ai Stresses on element adjacent to pile (a) Stresses on element adjacent to pile
Stress
cuo
Critical state
region
I
I
Stress
uo
4-
OCR=1
Ko
=
0.55
OCR=1
Ko
=
0.55
0
1 2
(bi Stress distribution around pile
10
r I II> I
100
I'/I'
1
I I I I I I > I
10
(bi Stress distribution around pile
I I I I I I
100
llr
Fig. 3. Theoretical stress state afterinstallation of pile in normally
consolidated soil (from Randolph er ak 1979b)
Fig. 5. Theoretical stress state after consolidation around pile in
normally consolidated soil (from Randolph er af, 1979b)
~
doi
3
lj
2-
(su/ovo~u
~+
1'
1
I
4 8
OCR
Fig. 4. Excess pore pressures generated
c/ose to a model jacked pile during instal-
latiOn (from Francescon, 1982)
model the precise detail of the soil move-
ment near the advancing pile tip. However,
it has been shown (Randolph et al, 1979a;
Steenfelt et al, 1981) that, over most of
the pile length, the radial soil displace-
ments due to installation of a displace-
ment pile are described well by idealising
the installation as the expansion of a cy-
lindrical cavity under conditions of plane
strain. This has enabled quantitative esti-
mates to be made of the stress changes
which occur in the soil around the pile
(Carter et al, 1979a; Randolph et al, 1979b) .
Idealisation of pile installation by cavity
expansion assumes that a cylindrical hole
is created, from zero initial radius, by radial
movement of the soil. The idealisation in-
volves a similar strain path to that follow-
ed in a pressuremeter test, and the total
stresses acting on the pile after cavity
expansion correspond to the limit pressure
achieved in such a test. Since the cavity
expansion starts from zero radius, the
shear strains in the soil are essentially
infinite adjacent to the pile wall, and are
such as to bring the soil to a state of fail-
ure out to a radius of 7-10 times the pile
radius, r,. Within this zone, the soil will
be remoulded to a greater or lesser extent.
The concepts embodied in critical state
soil mechanics indicate that soil sheared
to failure under undrained conditions will
tend to a unique effective stress state for
a given initial water content. Thus the
effective stress state in the failed zone of
soil around the pile will be uniform at any
particular depth. The total stress level will
vary with the radial distance from the pile
and the different changes in the total and
effective stress levels within any soil ele-
ment will give rise to excess pore pres-
sures. Typically, these may be as high as
4-6 times the in-situ shear strength close
to piles driven into normally or lightly
overconsolidated soils.
In the zone of soil sheared to critical
state conditions, where the effect've stress
conditions are uniform the shear strength
of the soil is fully mobilised in the hori-
zontal plane such that n-„'
n.fr
'
2c„,.
The equation of radial equilibrium then
implies that, within this zone, the radial
total stress, and hence the excess pore
pressures, should vary linearly with the
logarithm of the radial distance from the
pile axis, w'ith a gradient of 2c„,. Randolph
et al (1979b) have used the modified Cam
clay model (based on that described by
Roscoe 8f Burland, 1968) to provide quan-
titative estimates of the stresses after
cavity expansion. These are shown in Fig.
3 for soil with a friction angle of y'
30'.
5u
=
4c„,
—
3p'3)
where Ap's the change in mean effective
stress due to shearing the soil to a critical
state condition. For normally consolidated
soil, Ap'ill be negative and will typically
be of magnitude 1-1.5c„,.At increasing
values of overconsolidation ratio (OCR).
Ap'ill increase, becoming positive for
values of OCR greater than 2-3. Thus for
soil
'wet'f
critical (in the terminology
of Schofield 8s Wroth, 1968) the excess
pore pressures given by eqn. 3 will be
greater than
4cu,,
while for heavily over-
The numerical analyses showed little
variation of the size of the predicted ex-
cess pore pressures with the degree of
overconsolidation of the soil. However,
this result is questionable in view of the
greater tendency of heavily overconsolida-
ted soils to dilate. The excess pore pres-
sures arise partly from the increase in
mean total stress that accompanies cavity
expansion, an() partly as a result of chan-
ges in the mean effective stress as the soil
is sheared and remoulded. For normally or
lightly overconsolidated clay, the mean
effective stress will decrease during un-
drained shearing to failure, while for heav-
ily overconsolidated clay, there will be
an increase in the mean effective stress
(as the soil attempts to dilate). It may
therefore be expected that higher excess
pore pressures
w'll
be generated by pile
installation in lightly overconsolidated clay
than in heavily overconsolidated clay.
Randolph et al (1979b) suggest an ap-
proximate expression for estimating the
excess pore pressures generated adjacent
to a driven pile, which is
consolidated soils on the 'dry'ide of criti-
cal, <<p'ill be positive and excess pore
pressures of less than 4c„„are predicted.
This trend of decreasing excess pore pres-
sure with increasing OCR has been found
in model pile tests reported by
Frances-
con (1982). Fig. 4 shows some of his ex-
perimental results, where the excess pore
pressures generated at the pile shaft dur-
ing installation have been plotted against
the overconsolidation ratio of the clay.
The excess pore pressures have been nor-
malised by the initial undrained shear
strength,
c„„,
and also by the effective
overburden pressure, <re„'. Four test re-
sults are shown, corresponding to OCR of
1, 2, 4 and 8. The ratio
~u/ce,, decreases
from just under 5 at OCR
=
1 down to 2.6
at OCR
=
8. Correspondingly, the ratio
Au/<rr„'ncreases from unity at OCR
1, up to 2.7 at OCR =
8.
Although the simple model of cylindrical
cavity expansion has been justified on the
basis of the final radial movement of soil
around the pile, the stress changes will be
affected by the detailed strain path follow-
ed. A more detailed study of the stress
changes that occur has been made by
Levadoux (1980) by following the strain
path undergone by soil close to the ad-
vancing pile tip. The computations lead to
rather lower values of excess pore pres-
sures, particularly for heavily overconsoli-
dated soil.
2.2 Consolidation
The maximum hydraulic gradients exist-
ing immediately after installation of the
pile are primarily radial. At the pile wall,
the pile precludes straining of the soil, eit-
her along the pile length or around the
pile circumference, during consolidation.
Consequently, soil close to the pile under-
goes a form of one-dimensional consolida-
tion where the only non-zero strain is nor-
mal to the pile shaft (i.e. in the radial
direction). It is to be expected that the
radial effective stress will be the major
principal stress at the end of consolida-
tion. Finite element analyses have confirm-
ed this, and have shown that the vertical
and circumferential stresses at the end of
consolidation are approximately equal
(Randolph et al, 1979b). The finite element
analyses assumed that consolidation oc-
curs under plane strain conditions (with
no strain parallel to the pile axis, at any
radius). This assumption neglects the fact
that, in reality, there is a free ground sur-
face which may allow some straining of
the soil in the vertical direction.
For the assumption of plane strain, num-
erical predictions of the stress distribution
at the end of consolidation are as shown
in Fig, 5. Close to the pile, the radial and
vertical effective stresses are calculated
to be
rr5Ce<<<r3cec (4)
These values are reasonable for normal-
ly or lightly overconsolidated soil, where
c„„/<rr„'ay typically lie close to 0.3.
Thus the predicted final value of
<r
's
close to <rr„'. However, for heavily
overconsolidated clay, where
c„,/<rne may
be as high as 2 or
3, the predicted vertical
stre s appears unrealistically high, Free-
dom to strain in the vertical direction is
likely to give rise to rather lower values of
<rc
't
the end of the consolidation phase.
Of course, values of <r 'reater than the
effective overburden pressure may still de-
velop during consolidation, with vertical
equilibrium being maintained
by residual
shear stresses in the r:z plane. However,
during the working life of the pile, this
residual stress regime will alter and it
would seem reasonable to assume that o.
'ill
ultimately return to a value close to
the effective overburden pressure, n-„,'.
The consequences of this different be-
haviour, depending on the overconsolida-
tion ratio of the soil, are illustrated in Fig.
6 and 7. Fig. 6 shows stress paths obtained
by Randolph et al (1979b) for normally
consolidated clay. The paths are plotted
in q:p'nd V:p'pace, where p's
the
mean effective stress,
q
is the generalised
deviator stress given by qs =
0.5 [(<r
<ts')a + (<r.,' <r,')s +(o-s'-<T,')
'-'],
and
V is the specific volume of the soil. Start-
ing from point A, and ignoring ambient
pore pressures, the stress paths during
pile installation are shown by
AB'effec-
tive) and AB (total). During consolida-
tion, the effective and total stress states
converge, ending at point C. At this stage,
the water content of the soil has de-
creased and the soil strength will have in-
creased by 40
—
60%. This predicted be-
haviour is in good agreement with the
careful measurements of changes in soil
properties around piles driven in the field,
reported by Seed 8< Reese (1955). Wroth
et al (1979) report the results of finite
element analyses using the modified Cam-
clay model, showing how the increase in
shear strength varies with the consolida-
tion and strength properties of the soil.
Fig. 8 shows these results, with the
increase in shear strength,
c„„ /c„,,
(where c„represents the shear strength
of the soil after consolidation is complete)
being plotted against the Cam-clay para-
meter M. This parameter is related to the
friction angle <I>
'or
triaxial compression
by M
=
6sin<i<'/(3
—
sin<~').
As discussed earlier, the stress paths for
heavily overconsolidated soil may differ
from those for normally consolidated soil,
especially where straining is permitted in
the vertical direction during consolidation.
Fig. 7 indicates possible stress paths for
heavily overconsolidated soil. During in-
stallation, lower excess pore pressures are
generated, and the final effective stress
level after consolidation, is controlled by
the condition that <t-.'eturns to close to
<r„„'. The increase in the strength of the
soil during consolidation will be less than
for the case of a pile driven into normally
consolidated soil.
As consolidation proceeds, the effective
stresses adjacent to the pile will increase
Critical
state line
Total stress
paths
+B
Total stress
paths
P,P
P,P
Normal
consolidation line
Critical
state line
ormal
nsolidation line
Critical
state I <ne
Fig. 6. Stress paths for normally consolidated soil Fig. 7. Stress paths for heavily overconsolidated soil
P
and the pile capacity will show a corres-
ponding increase. The variation of pile ca-
pacity with time after installation has been
measured from field tests by a number of
workers (Seed & Reese, 1955; Eide et a(,
1961; Thorburn & Rigden, 1980). The para-
meters x and
P
deduced from pile load
tests will therefore be influenced by when
the test is carried out. In the remainder
of this Paper, the subscript oc will be used
to indicate the long-term values of these
parameters after all consolidation is com-
plete. Since the soil strength increases by
some 50% during the consolidation pro-
cess, it is not unreasonable to expect a
similar increase in pile capacity, and thus
in x and
P,
between a load test performed
immediately following pile installation, and
one performed after full consolidation. In
the past, many load tests on piles have
been conducted relatively soon after in-
stallation, and the resulting deduced val-
ues of x and
P
may thus be unrepresenta-
tive of the long-term values.
An appropriate dimensionless group
which determines the rate of consolidation
around a driven pile is T
=
c„t/r,-', where
c>,
is the coefficient of consolidation for
horizontal drainage, t is the elapsed time
and r, is the radius of the pile. Randolph
et a( (1979b) show values of T required
for 50 and 90% dissipation of excess pore
pressures. Typically, values for T„are
about 30 for normally or lightly overcon-
solidated soil, decreasing to about 10 for
more heavily overconsolidated soil where
the zone of excess pore pressures is less
widespread. The coefficient of consolida-
tion, c„,
for a particular soi'I will be strong-
ly affected by
whether the soil is following
a path of virgin consolidation, or whether
it is following a swelling or re-loading path.
In finite element analyses performed using
the modified Cam-clay model, which auto
matically takes account of which type of
path is being followed, Randolph et al
(1979b) adopt an alternative dimension-
less time factor, based on the permea-
bility of the soil. The analyses showed
that, although soil immediately around the
pile follows a virgin consolidation path,
most of the soil (beyond
2-3 pile radii)
unloads in shear during consolidation. The
consolidation times are thus governed by
a coefficient of consolidation appropriate
for swelling paths.
As an example of the time needed for
90% dissipation of excess pore pressures,
consider the case of a 0.5m diameter pile
driven into soil with c„= 10
c
ma/s. Tak-
ing 7e„= 30, the time for 90% dissipa-
tion is calculated to be about 20 days.
When determining values of skin friction
for use in design, it is therefore important
to allow a suitable time to elapse be-
tween pile installation and conducting a
load test.
2.3 Pile loading
When a pile is loaded axially in com-
pression, load is transferred to the soil by
shear stresses acting at the pile shaft, and
initially horizontal layers of soil are de-
formed as shown in Fig. 9(a). The strain
path undergone by elements of soil close
to the pile shaft is similar to that imposed
on soil in a simple shear test. Comparing
Fig. 9(b) with Fig. 9(a), there is a direct
correspondence between the
y
and x dir-
ections of the simple shear test, and the
r and z directions of the pile case. Ran-
dolph & Wroth (1981) have explored this
correspondence with a view to estimating
the ultimate skin friction between pile and
soil from given effective stress conditions
in the soil immediately prior to loading the
pile.
One of the important differences bet-
ween an axially loaded pile and the sim-
ple shear test is that, in the former, the
possibility of slip occurring between the
pile and the adjacent soil element must
be considered. This possibility leads to
two alternative modes of failure of the
pile. In the first of these, the soil con-
tinuum adjacent to the pile is brought to
a failure state without slip. The current
Mohr's circle for stress will be tangential
to the Mohr-Coulomb failure criterion as
shown in Fig. 10(a). As the pile plunges
to failure, the directions of principal stress
axes will rotate until the angle of friction
mobilised on the pile shaft reaches its
maximum value (here taken equal to 11'—
the angle of internal friction for the soil).
These conditions at failure have been dis-
cussed by Parry & Swain (1977).
An alternative mode of failure is pos-
sible where the soil possesses some co-
hesion. The bond between the pile and
soil may be considered purely frictional
and so it becomes possible for the pile
to plunge to failure even though the maxi-
mum shear stress mobilised is consider-
ably less than the current strength of the
soil continuum. Fig. 10(b) shows this sit-
uation; rupture between the pile and the
soil has intervened before the full soil
strength has been mobilised, at a valuq of
skin friction given by
o-,'an(i
(5)
This type of failure forms the basis of the
approach to predicting pile capacity des-
cribed by Burland (1973).
It is possible to use concepts of critical
state soil mechanics to estimate the skin
friction in the first type of failure discussed
above. Since the soil continuum has been
brought locally to a state of failure, the cur-
rent mean effective stress may be taken as
that corresponding to the critical state for
the current water content of the soil. Since
the mode of deformation (simple shear) is
essentially one of plane strain, the centre
of the Mohr's circle of stress may be
assumed approximately equal to this mean
effective stress, p('. (This is equivalent to
taking the intermediate principal stress,
<r,', as equal to the average of the major
and minor principal stresses). From Fig.
10(a), the skin friction on the pile shaft
is then given by
t.,
=
p('ing 'osy
'
(6)
whence the skin friction, -,„, may be ex-
pressed in terms of the current strength
as
3
—
SI noh
cosy (c„)„... (8)
3
Randolph & Wroth (1981) have shown
(see Fig. 11) that the above expression
gives a reasonably good fit to experimen-
tal data of the ratio between peak stren-
gths measured in simple shear and in
triaxial compression. The data are from
tests on nine natural clays and the figure
The value of p('ay be related to the
deviator stress at failure,
q(,
and hence to
the current undrained shear strength in
triaxial compression, (c,)„by
q(
3
—
sinrjr
','
— =
(c„)„"(7)
M 3siny
'
oo
VD
1.5-
4
—
~— —
a
— — —-
d
0.5—
Legend
~ London clay
~ Boston Blue clay
Weald clay
Kaolin
o Gault clay
/
/
v
fr ~aa
(a) Axially loaded pile
p
— /-
(b) Simple shear test
Fig. 9. Analogy between shearing of soil
around an axially loaded pile and soilin a
simple shear test
(a) Failure of soil continuum for normally consolidated
soil
Failure e
for soil c
criterion
il interface
Current
stress circle
(b) Rupture at pile-soil interface prior to failure of soil
continuum for heavily overconsolidated soil
Fig. 10. Possible modes of failure at pile-soil
i nterface
0
0.5 0.75 1.25 1.5
M
Fig. 8. Gain in soil strength during con-
so(idation for different soils
(from Wroth et a(, 1979)
Page 5


1. Introduction
THERE ARE many areas of geotechnical
engineering where design is based largely
on empirical methods that have evolved
from previous experience. Such approa-
ches may work well in familiar situations,
but there will inevitably be uncertainty
when novel structures or new types of
soil deposits are encountered. Pile founda-
tions have been designed more or less
successfully, using an empirical approach,
for a large number of years. However, the
advent of the offshore oil industry has
necessitated the design of piles with ca-
pacity of an order of magnitude higher
than onshore piles, and in soils which are
in many instances different from onshore
deposits. This has provided considerable
impetus to improving the understanding
of how driven piles develop their capacity,
in order to provide a more rational basis
for pile design.
Nearly all the working load capacity of
driven piles comes from the shaft, and this
Paper is restricted to discussion of devel-
opments in the prediction of the shaft ca-
pacity of piles driven into clay. Until rela-
tively recently, the prevalent design phil-
osophy was that some bond or adhesion
existed between the pile shaft and the soil.
It was natural to correlate the strength of
this bond with the undrained shear stren-
gth of the soil. This led to what is com-
monly referred to as the 'total stress met-
hod'or
predicting pile capacity. The limit-
ing skin friction, —,„,on the pile shaft is
expressed as a proportion of the in-situ
shear strength of the soil,
cu,
as
<a
=
"cue
Experience obtained from onshore pile
tests has enabled the parameter x to be
correlated with the type and strength of
the soil (e.g. Tomlinson 1980).
The first advances from this standpoint
were due to Chandler (1968) and Burland
(1973) who argued that soil behaviour is
controlled by effective stresses and that
the shaft capacity of a pile should be view-
ed in the light of a limiting frictional stress
developed between the pile and the soil.
The new 'effective stress'pproach sought
to relate the skin friction to the in-situ
effective stress state. Although it was rec-
ognised that the limiting
sk'n
friction would
depend primarily on the horizontal effec-
tive stress, this quantity is difficult to esti-
mate with confidence. The most accurately
determinable effective stress in the soil is
the effective overburden pressure,
rre,
'.
Consequently, the effect've stress a p-
proach has led to the skin friction being
expressed as some fraction,
P,
of the effec-
«University of Cambridge
liuniversity of Oxford
This Paper will form the bas's of a presentation
by the authors to an Informal Discussion meet-
ing of the British Geotechnical Society on Dec. t
this year at the Institution of Civil Engineers,
Westminster, London.
tive overburden pressure according to
Ta P(rte (2)
The parameter
P
thus reflects not only
the friction angle between the pile and the
soil, but also the ratio of the horizontal
and vertical effective stresses. Burland
(1973) suggested that this ratio could be
taken as equal to the at-rest earth pres-
sure coefficient,
K,, acknowledging that
this would be conservative for driven piles,
where the act of installing the pile may
locally increase the horizontal stress in the
soil.
Although derived from different stand-
points, for the designer there is no particu-
lar merit to the expression in eqn. 2 com-
pared with that in eqn. 1. In both cases,
the skin friction is directly and simply re-
lated to measurable properties of the soil
which reflect its strength prior to install-
ing a pile. But of course, the advantage
of the effective stress approach is that it
opens the door to considering, in detail,
precisely what changes in total and effec-
tive stresses take place due to the irstal-
lation of the pile and due to its subse-
quent loading.
This Paper outlines some of the more
recent developments that have occurred
in the effective stress approach to pile de-
sign. In particular, attention is concentra-
ted on three main events in the history of
a driven pile. These are
(i) pile installation
(ii
)
consolidation, as excess pore pres-
sures generated in (i) dissipate, and
(iii) pile loading.
Much of the work described herein was
instigated by the ESACC project, organised
by Amoco Production Company and fund-
ed by 17 assorted organisations. A sum-
mary of this project may be found in Kraft
et al (1980) and a review of the effective
stress methods of design that emerged
from the project is given by Kraft (1982).
One of the chief limitations to formu-
lating new design rules has been the pauc-
ity of reliable data on the shaft capacity
of piles, particularly in the types of soil
encountered offshore. In the absence of
such data, some of the discussion in the
present Paper is necessarily of a specula-
tive nature; the Paper should be viewed
as an attempt to draw preliminary design
guidelines from research which is still con-
tinuing and has some way to go before
more definitive statements may be made.
2. Changes in stresses
In discussing the stress changes which
occur in the soil, it is convenient to adopt
cylindrical polar co-ordinates r, 6'nd
z,
with the z axis directed down the axis of
the pile as shown in Fig. 1. Discussion will
be confined to cylindrical piles but the
principles developed are also applicable to
piles of other cross-sectional shapes or to
tapered piles. Prior to pile installation, the
stresses acting on a typical element of
soil are the 'at rest'tresses 0.,' —
<r„o'nd
rr
'
tran
'
K rr +', as shown in Fig.
1. In accordance with common practice, it
is assumed that the horizontal stresses
are the same in any direction. The changes
in the stresses during and after pile instal-
lation are now discussed.
2.1 Pile installation
The process of installing a displacement
pile into clay is one which is not easily
modelled analytically. Fig. 2 shows sche-
matically the main soil movements which
occur. The most common idealisation of
these movements has been that of cylin-
drical cavity expansion (e.g. Soderberg,
1962). The idealisation neglects local ef-
fects such as heave of the ground surface
around the pile and does not
attempt
to
<')
~0
r
Pile
Soil element
Fig. 1. Co-ordinate system and initial
stress state in the soil
Pile~— Surface
heave
~)Z,
I
I
Zone of intensely~',
remoulded so
"l
Radial movement
of soil
Spherical
pressure bulb
Fig. 2. Soil movements due to pile instal-
lationn
o,'-
1.5cuo oz 3cuo
or 2.5cuo u,
—
5cu
(ai Stresses on element adjacent to pile (a) Stresses on element adjacent to pile
Stress
cuo
Critical state
region
I
I
Stress
uo
4-
OCR=1
Ko
=
0.55
OCR=1
Ko
=
0.55
0
1 2
(bi Stress distribution around pile
10
r I II> I
100
I'/I'
1
I I I I I I > I
10
(bi Stress distribution around pile
I I I I I I
100
llr
Fig. 3. Theoretical stress state afterinstallation of pile in normally
consolidated soil (from Randolph er ak 1979b)
Fig. 5. Theoretical stress state after consolidation around pile in
normally consolidated soil (from Randolph er af, 1979b)
~
doi
3
lj
2-
(su/ovo~u
~+
1'
1
I
4 8
OCR
Fig. 4. Excess pore pressures generated
c/ose to a model jacked pile during instal-
latiOn (from Francescon, 1982)
model the precise detail of the soil move-
ment near the advancing pile tip. However,
it has been shown (Randolph et al, 1979a;
Steenfelt et al, 1981) that, over most of
the pile length, the radial soil displace-
ments due to installation of a displace-
ment pile are described well by idealising
the installation as the expansion of a cy-
lindrical cavity under conditions of plane
strain. This has enabled quantitative esti-
mates to be made of the stress changes
which occur in the soil around the pile
(Carter et al, 1979a; Randolph et al, 1979b) .
Idealisation of pile installation by cavity
expansion assumes that a cylindrical hole
is created, from zero initial radius, by radial
movement of the soil. The idealisation in-
volves a similar strain path to that follow-
ed in a pressuremeter test, and the total
stresses acting on the pile after cavity
expansion correspond to the limit pressure
achieved in such a test. Since the cavity
expansion starts from zero radius, the
shear strains in the soil are essentially
infinite adjacent to the pile wall, and are
such as to bring the soil to a state of fail-
ure out to a radius of 7-10 times the pile
radius, r,. Within this zone, the soil will
be remoulded to a greater or lesser extent.
The concepts embodied in critical state
soil mechanics indicate that soil sheared
to failure under undrained conditions will
tend to a unique effective stress state for
a given initial water content. Thus the
effective stress state in the failed zone of
soil around the pile will be uniform at any
particular depth. The total stress level will
vary with the radial distance from the pile
and the different changes in the total and
effective stress levels within any soil ele-
ment will give rise to excess pore pres-
sures. Typically, these may be as high as
4-6 times the in-situ shear strength close
to piles driven into normally or lightly
overconsolidated soils.
In the zone of soil sheared to critical
state conditions, where the effect've stress
conditions are uniform the shear strength
of the soil is fully mobilised in the hori-
zontal plane such that n-„'
n.fr
'
2c„,.
The equation of radial equilibrium then
implies that, within this zone, the radial
total stress, and hence the excess pore
pressures, should vary linearly with the
logarithm of the radial distance from the
pile axis, w'ith a gradient of 2c„,. Randolph
et al (1979b) have used the modified Cam
clay model (based on that described by
Roscoe 8f Burland, 1968) to provide quan-
titative estimates of the stresses after
cavity expansion. These are shown in Fig.
3 for soil with a friction angle of y'
30'.
5u
=
4c„,
—
3p'3)
where Ap's the change in mean effective
stress due to shearing the soil to a critical
state condition. For normally consolidated
soil, Ap'ill be negative and will typically
be of magnitude 1-1.5c„,.At increasing
values of overconsolidation ratio (OCR).
Ap'ill increase, becoming positive for
values of OCR greater than 2-3. Thus for
soil
'wet'f
critical (in the terminology
of Schofield 8s Wroth, 1968) the excess
pore pressures given by eqn. 3 will be
greater than
4cu,,
while for heavily over-
The numerical analyses showed little
variation of the size of the predicted ex-
cess pore pressures with the degree of
overconsolidation of the soil. However,
this result is questionable in view of the
greater tendency of heavily overconsolida-
ted soils to dilate. The excess pore pres-
sures arise partly from the increase in
mean total stress that accompanies cavity
expansion, an() partly as a result of chan-
ges in the mean effective stress as the soil
is sheared and remoulded. For normally or
lightly overconsolidated clay, the mean
effective stress will decrease during un-
drained shearing to failure, while for heav-
ily overconsolidated clay, there will be
an increase in the mean effective stress
(as the soil attempts to dilate). It may
therefore be expected that higher excess
pore pressures
w'll
be generated by pile
installation in lightly overconsolidated clay
than in heavily overconsolidated clay.
Randolph et al (1979b) suggest an ap-
proximate expression for estimating the
excess pore pressures generated adjacent
to a driven pile, which is
consolidated soils on the 'dry'ide of criti-
cal, <<p'ill be positive and excess pore
pressures of less than 4c„„are predicted.
This trend of decreasing excess pore pres-
sure with increasing OCR has been found
in model pile tests reported by
Frances-
con (1982). Fig. 4 shows some of his ex-
perimental results, where the excess pore
pressures generated at the pile shaft dur-
ing installation have been plotted against
the overconsolidation ratio of the clay.
The excess pore pressures have been nor-
malised by the initial undrained shear
strength,
c„„,
and also by the effective
overburden pressure, <re„'. Four test re-
sults are shown, corresponding to OCR of
1, 2, 4 and 8. The ratio
~u/ce,, decreases
from just under 5 at OCR
=
1 down to 2.6
at OCR
=
8. Correspondingly, the ratio
Au/<rr„'ncreases from unity at OCR
1, up to 2.7 at OCR =
8.
Although the simple model of cylindrical
cavity expansion has been justified on the
basis of the final radial movement of soil
around the pile, the stress changes will be
affected by the detailed strain path follow-
ed. A more detailed study of the stress
changes that occur has been made by
Levadoux (1980) by following the strain
path undergone by soil close to the ad-
vancing pile tip. The computations lead to
rather lower values of excess pore pres-
sures, particularly for heavily overconsoli-
dated soil.
2.2 Consolidation
The maximum hydraulic gradients exist-
ing immediately after installation of the
pile are primarily radial. At the pile wall,
the pile precludes straining of the soil, eit-
her along the pile length or around the
pile circumference, during consolidation.
Consequently, soil close to the pile under-
goes a form of one-dimensional consolida-
tion where the only non-zero strain is nor-
mal to the pile shaft (i.e. in the radial
direction). It is to be expected that the
radial effective stress will be the major
principal stress at the end of consolida-
tion. Finite element analyses have confirm-
ed this, and have shown that the vertical
and circumferential stresses at the end of
consolidation are approximately equal
(Randolph et al, 1979b). The finite element
analyses assumed that consolidation oc-
curs under plane strain conditions (with
no strain parallel to the pile axis, at any
radius). This assumption neglects the fact
that, in reality, there is a free ground sur-
face which may allow some straining of
the soil in the vertical direction.
For the assumption of plane strain, num-
erical predictions of the stress distribution
at the end of consolidation are as shown
in Fig, 5. Close to the pile, the radial and
vertical effective stresses are calculated
to be
rr5Ce<<<r3cec (4)
These values are reasonable for normal-
ly or lightly overconsolidated soil, where
c„„/<rr„'ay typically lie close to 0.3.
Thus the predicted final value of
<r
's
close to <rr„'. However, for heavily
overconsolidated clay, where
c„,/<rne may
be as high as 2 or
3, the predicted vertical
stre s appears unrealistically high, Free-
dom to strain in the vertical direction is
likely to give rise to rather lower values of
<rc
't
the end of the consolidation phase.
Of course, values of <r 'reater than the
effective overburden pressure may still de-
velop during consolidation, with vertical
equilibrium being maintained
by residual
shear stresses in the r:z plane. However,
during the working life of the pile, this
residual stress regime will alter and it
would seem reasonable to assume that o.
'ill
ultimately return to a value close to
the effective overburden pressure, n-„,'.
The consequences of this different be-
haviour, depending on the overconsolida-
tion ratio of the soil, are illustrated in Fig.
6 and 7. Fig. 6 shows stress paths obtained
by Randolph et al (1979b) for normally
consolidated clay. The paths are plotted
in q:p'nd V:p'pace, where p's
the
mean effective stress,
q
is the generalised
deviator stress given by qs =
0.5 [(<r
<ts')a + (<r.,' <r,')s +(o-s'-<T,')
'-'],
and
V is the specific volume of the soil. Start-
ing from point A, and ignoring ambient
pore pressures, the stress paths during
pile installation are shown by
AB'effec-
tive) and AB (total). During consolida-
tion, the effective and total stress states
converge, ending at point C. At this stage,
the water content of the soil has de-
creased and the soil strength will have in-
creased by 40
—
60%. This predicted be-
haviour is in good agreement with the
careful measurements of changes in soil
properties around piles driven in the field,
reported by Seed 8< Reese (1955). Wroth
et al (1979) report the results of finite
element analyses using the modified Cam-
clay model, showing how the increase in
shear strength varies with the consolida-
tion and strength properties of the soil.
Fig. 8 shows these results, with the
increase in shear strength,
c„„ /c„,,
(where c„represents the shear strength
of the soil after consolidation is complete)
being plotted against the Cam-clay para-
meter M. This parameter is related to the
friction angle <I>
'or
triaxial compression
by M
=
6sin<i<'/(3
—
sin<~').
As discussed earlier, the stress paths for
heavily overconsolidated soil may differ
from those for normally consolidated soil,
especially where straining is permitted in
the vertical direction during consolidation.
Fig. 7 indicates possible stress paths for
heavily overconsolidated soil. During in-
stallation, lower excess pore pressures are
generated, and the final effective stress
level after consolidation, is controlled by
the condition that <t-.'eturns to close to
<r„„'. The increase in the strength of the
soil during consolidation will be less than
for the case of a pile driven into normally
consolidated soil.
As consolidation proceeds, the effective
stresses adjacent to the pile will increase
Critical
state line
Total stress
paths
+B
Total stress
paths
P,P
P,P
Normal
consolidation line
Critical
state line
ormal
nsolidation line
Critical
state I <ne
Fig. 6. Stress paths for normally consolidated soil Fig. 7. Stress paths for heavily overconsolidated soil
P
and the pile capacity will show a corres-
ponding increase. The variation of pile ca-
pacity with time after installation has been
measured from field tests by a number of
workers (Seed & Reese, 1955; Eide et a(,
1961; Thorburn & Rigden, 1980). The para-
meters x and
P
deduced from pile load
tests will therefore be influenced by when
the test is carried out. In the remainder
of this Paper, the subscript oc will be used
to indicate the long-term values of these
parameters after all consolidation is com-
plete. Since the soil strength increases by
some 50% during the consolidation pro-
cess, it is not unreasonable to expect a
similar increase in pile capacity, and thus
in x and
P,
between a load test performed
immediately following pile installation, and
one performed after full consolidation. In
the past, many load tests on piles have
been conducted relatively soon after in-
stallation, and the resulting deduced val-
ues of x and
P
may thus be unrepresenta-
tive of the long-term values.
An appropriate dimensionless group
which determines the rate of consolidation
around a driven pile is T
=
c„t/r,-', where
c>,
is the coefficient of consolidation for
horizontal drainage, t is the elapsed time
and r, is the radius of the pile. Randolph
et a( (1979b) show values of T required
for 50 and 90% dissipation of excess pore
pressures. Typically, values for T„are
about 30 for normally or lightly overcon-
solidated soil, decreasing to about 10 for
more heavily overconsolidated soil where
the zone of excess pore pressures is less
widespread. The coefficient of consolida-
tion, c„,
for a particular soi'I will be strong-
ly affected by
whether the soil is following
a path of virgin consolidation, or whether
it is following a swelling or re-loading path.
In finite element analyses performed using
the modified Cam-clay model, which auto
matically takes account of which type of
path is being followed, Randolph et al
(1979b) adopt an alternative dimension-
less time factor, based on the permea-
bility of the soil. The analyses showed
that, although soil immediately around the
pile follows a virgin consolidation path,
most of the soil (beyond
2-3 pile radii)
unloads in shear during consolidation. The
consolidation times are thus governed by
a coefficient of consolidation appropriate
for swelling paths.
As an example of the time needed for
90% dissipation of excess pore pressures,
consider the case of a 0.5m diameter pile
driven into soil with c„= 10
c
ma/s. Tak-
ing 7e„= 30, the time for 90% dissipa-
tion is calculated to be about 20 days.
When determining values of skin friction
for use in design, it is therefore important
to allow a suitable time to elapse be-
tween pile installation and conducting a
load test.
2.3 Pile loading
When a pile is loaded axially in com-
pression, load is transferred to the soil by
shear stresses acting at the pile shaft, and
initially horizontal layers of soil are de-
formed as shown in Fig. 9(a). The strain
path undergone by elements of soil close
to the pile shaft is similar to that imposed
on soil in a simple shear test. Comparing
Fig. 9(b) with Fig. 9(a), there is a direct
correspondence between the
y
and x dir-
ections of the simple shear test, and the
r and z directions of the pile case. Ran-
dolph & Wroth (1981) have explored this
correspondence with a view to estimating
the ultimate skin friction between pile and
soil from given effective stress conditions
in the soil immediately prior to loading the
pile.
One of the important differences bet-
ween an axially loaded pile and the sim-
ple shear test is that, in the former, the
possibility of slip occurring between the
pile and the adjacent soil element must
be considered. This possibility leads to
two alternative modes of failure of the
pile. In the first of these, the soil con-
tinuum adjacent to the pile is brought to
a failure state without slip. The current
Mohr's circle for stress will be tangential
to the Mohr-Coulomb failure criterion as
shown in Fig. 10(a). As the pile plunges
to failure, the directions of principal stress
axes will rotate until the angle of friction
mobilised on the pile shaft reaches its
maximum value (here taken equal to 11'—
the angle of internal friction for the soil).
These conditions at failure have been dis-
cussed by Parry & Swain (1977).
An alternative mode of failure is pos-
sible where the soil possesses some co-
hesion. The bond between the pile and
soil may be considered purely frictional
and so it becomes possible for the pile
to plunge to failure even though the maxi-
mum shear stress mobilised is consider-
ably less than the current strength of the
soil continuum. Fig. 10(b) shows this sit-
uation; rupture between the pile and the
soil has intervened before the full soil
strength has been mobilised, at a valuq of
skin friction given by
o-,'an(i
(5)
This type of failure forms the basis of the
approach to predicting pile capacity des-
cribed by Burland (1973).
It is possible to use concepts of critical
state soil mechanics to estimate the skin
friction in the first type of failure discussed
above. Since the soil continuum has been
brought locally to a state of failure, the cur-
rent mean effective stress may be taken as
that corresponding to the critical state for
the current water content of the soil. Since
the mode of deformation (simple shear) is
essentially one of plane strain, the centre
of the Mohr's circle of stress may be
assumed approximately equal to this mean
effective stress, p('. (This is equivalent to
taking the intermediate principal stress,
<r,', as equal to the average of the major
and minor principal stresses). From Fig.
10(a), the skin friction on the pile shaft
is then given by
t.,
=
p('ing 'osy
'
(6)
whence the skin friction, -,„, may be ex-
pressed in terms of the current strength
as
3
—
SI noh
cosy (c„)„... (8)
3
Randolph & Wroth (1981) have shown
(see Fig. 11) that the above expression
gives a reasonably good fit to experimen-
tal data of the ratio between peak stren-
gths measured in simple shear and in
triaxial compression. The data are from
tests on nine natural clays and the figure
The value of p('ay be related to the
deviator stress at failure,
q(,
and hence to
the current undrained shear strength in
triaxial compression, (c,)„by
q(
3
—
sinrjr
','
— =
(c„)„"(7)
M 3siny
'
oo
VD
1.5-
4
—
~— —
a
— — —-
d
0.5—
Legend
~ London clay
~ Boston Blue clay
Weald clay
Kaolin
o Gault clay
/
/
v
fr ~aa
(a) Axially loaded pile
p
— /-
(b) Simple shear test
Fig. 9. Analogy between shearing of soil
around an axially loaded pile and soilin a
simple shear test
(a) Failure of soil continuum for normally consolidated
soil
Failure e
for soil c
criterion
il interface
Current
stress circle
(b) Rupture at pile-soil interface prior to failure of soil
continuum for heavily overconsolidated soil
Fig. 10. Possible modes of failure at pile-soil
i nterface
0
0.5 0.75 1.25 1.5
M
Fig. 8. Gain in soil strength during con-
so(idation for different soils
(from Wroth et a(, 1979)
1.0-
Tf
«u>C»
0.8-
Ladd sr E
PI 30
—
80
122'
fs'6-
'
3
0.4-
4
5
6
02-
'
9
Boston Blue clay
Maine Organic clay
Bangkok clay
Atchafalya clay
Portland Maine clay
Portsmouth Sensitive clay
Connecticut Valley Varved clay
Kaolin
Orammen clay
1 ~
5
Ladd & Edgers
Pl 10
—
25%
(30 < fo'< 35')
0.4
0.2—
1 I I I
0 5 10 15 20 25 30 35
4
10
Fig. 11 (above). Ratio of strength in simple shear test to strength
in triaxial compression (from Randolph fk wroth, 1981)
Fig. 12 (right). Suggested design curves fer variation of x„and
P„with
ce,/o-„'s
fl=-
ouo
0
0.2
1.5
0.4 0.8 1.6 3.2
cuoforo
Ko
tan fS
—
fs'=
30'8'lso
shows recommendations by Ladd
8t Edgers (1972) for the ratio
T,/(c„) „de-
pending on the plasticity index (Pl) of
the soil.
2.4 Summary
In summarising the points made in this
section, it is helpful to consider the two
extremes of a normally consolidated clay
and a heavily overconsolidated clay.
0.5
0
0.2 0.4 0.8 1.6
—
22
—
20'.2
(a) Normally consolidated c/ay
During pile installation, the effective
stresses reduce and the total stresses in-
crease, giving excess pore pressures next
to the pile of around 4-5ceo (see Figs. 3
and 4). Dissipation of these excess pore
pressures causes a decrease in water con-
tent close to the pile and a resulting in-
crease in strength of 40-60%. The final
effective stress state in the soil close to
the pile may be estimated as
~z' ~vo''~r' ~vo'/(
o)ne
.
( )
where
(K,)
„,
is the value of
K,
for a
normally consolidated deposit of the soil.
For typical values of
c~/fr„o
'or
soft clay,
these estimates will correspond closely to
those given in eqn. 4 and shown in Fig.
5. When the pile is loaded to failure, the
effective stresses will reduce, (see Potts 8t
Martins, 1982) as in any undrained shear
test on normally consolidated clay, and the
final skin friction may be estimated using
eqn. 8, where now
(c„)„represents the
current shear strength as measured in triax
ial compression. For y' 30', this equa-
tion gives
T,/(c„)
sc
—
—
0.72. Since the
strength of the soil is now some 50%
greater than previously, the resulting ratio
of skin friction, —... to the original shear
strength
c„o
is x„=
T,/cuo
=
0.72 x
1.5 =
1.08. Thus, as indicated by Ran-
dolph 8t Wroth (1981), the final predicted
value for normally consolidated clay
is just over unity. This conforms with ex-
perimental evidence from piles driven into
soft clay.
(b) Heavily overconsolidated clay
As the soil is sheared during pile instal-
lation, the total stress increase will be
accompanied by an increase in the effec-
tive stress level as the soil attempts to
dilate. The resulting excess pore pressures
will be smaller than for the case of a pile
driven into normally consolidated clay, but
they may still be as high as 2-3 times the
in-situ undrained strength, c„,. Dissipation
of these pore pressures may be accom-
panied by little, if any, increase in the
effective stress level. Freedom of the soil
to strain in the vertical direction may re-
strict the final vertical effective stress to
close to rr„o'. When the pile is loaded to
failure, premature rupture may occur at the
pile-soil interface (where the cohesion is
assumed zero) before the soil continuum
reaches failure, as shown in Fig. 10(b).
Meyerhof (1976) has analysed the re-
sults of a number of load tests on piles
driven into heavily overconsolidated clay.
Assuming that the friction angle between
the pile and soil is the same as that inter
nally in the soil (i.e. Fi
= y'n
eqn. 5), he
shows that the normal effective stress,
o-„'cting
on the pile at failure, lies between
1 and 2 times the best estimate of the
original horizontal effective stress,
with an average value of 1.5rr„o'. Thus the
limiting skin friction may be expressed as
—,,
=
1.5K, frvo'anIg'
10)
The corresponding expression for the para
meter
P
then becomes
3. Design parameters for unit skin
friction values
There are still many areas of uncertainty
in the qualitative model of soil response
outlined in the previous section. An ob-
vious one concerns the assumption that,
P
=
rs/rr„'
—
1.5Ko tanIg'..
(11)
This expression for
p,
although deduced
from actual pile load tests, must be viewed
in the light, firstly, of the assumption that
5
= y'nd, secondly, of possible inaccur-
acies in the estimates of
Ko for the sites
in question. However, it provides a tenta-
tive basis for estimating the shaft capac-
ity of piles driven into heavily overcon-
solidated clay.
cuo/oro
close to the pile, rr,'ends to reduce to-
wards er„,'fter consolidation. This assump-
tion seems reasonable in the long term.
However, it is possible that higher effective
stresses may be locked in the soil as
consolidation proceeds, together with a
relatively complex pattern of residual
stresses, as discussed by Clegg (1981).
Subsequently, during the working life of
the pile, the pattern of residual stresses
will change, and strains within the soil
mass may allow the vertical effective
stress, rr,', to reduce to a value close to
Such a pattern of stress changes
would still leave the soil close to the pile
at a lower water content (and increased
strength) even for piles driven into heavily
overconsolidated clay. It would also lead
to the possibility of high pile capacities
being recorded in the medium term (i.e.
after consolidation, but while the high effec-
tive szsesses are still locked in). However,
as a design philosophy, it would seem pru-
dent to make the assumption that the verti-
cal effective stress at any depth returns
to close to the original effective overburden
pressure.
Any natural soil deposit will show a
variation of the overconsolidation ratio
with depth. It is a feature of any effective
stress approach to predicting pile capacity,
that parameters such as x or
P
will vary
with OCR, and hence also with depth for
a given location. Such a variation with
depth is already implied in some design
formulae, notably that using the A, para-
meter (Vijayvergiya 8t Focht, 1972) where
the skin friction is estimated as
Ts
=
A(2c„o + fr„,') (12)
where
cn,
and rr„'re the average values
of initial undrained shear strength and ef-
fective overburden pressure, over the
depth of em'bedment of the pile. Tradition-
ally, this design approach uses a single
global A value for any given pile at a par-